Early Identification of Delirium in Intensive Care Unit Patients: Improving the Quality of Care.

This was a quality improvement pilot in one 16-bed mixed medical ICU, so it may not apply to other units. The design was observational pretest-posttest, and several education and screening components were delivered together, so it cannot show which component led to the medication decrease. Nurse survey response was low: 24 of 59 nurses at baseline and 15 of 59 after implementation, and the two surveys were not paired, so changes may reflect who responded rather than all nurses. CAM-ICU documentation was 83% at six weeks, but many assessments were missing, many were unable to assess, and the EHR allowed opting out; nurses may have skipped the RASS. Patients were not identified as having delirium before implementation, so risk groups were based on admitting diagnoses, and some already had delirium symptoms or comorbidities. Opioid and benzodiazepine dose and frequency were not considered, and extreme outliers affected analyses. The CAM-ICU flow sheet remained active for some patients transferred to step-down units, so untrained nurses documented it and tracking was difficult. This was approved as a nonresearch quality improvement project, not an institutional review board research study.

Do not conclude that CAM-ICU screening alone reduced opioid and benzodiazepine use or shortened ICU stays. The intervention also included online education, hands-on training, pocket cards, EHR CAM-ICU documentation and spot checks, and length of stay did not significantly decrease.